I attached the to exe in the bin\debug dir and when I profiled it, I did not have any performance issues that I could find. If I run my app from within VS, then the performance issues show up. I didn't really consider it a problem, rather the way dotTrace works, but perhaps I'm not using dotTrace correctly. Any suggestions?
Right now, I'm considering ripping out AR and just using NHibernate directly. Not sure how much time I want to put into figuring this issue out. scott On Aug 21, 6:54 pm, John Simons <[email protected]> wrote: > Scott, > > What exacly is the problem you are having with profiling you app with > dotTrace? > Profiling an app that is being debugged by VS or or not is the same, there > shouldn't be any extra steps. > > BTW, I'm also experiencing the performance issue you have reported, and I do > not have a work around. I've attached my profiler screenshots of where it > slows down. As you can see the CLR method CreateType is the culprid! > > Cheers > John > > ________________________________ > From: scottieb <[email protected]> > To: Castle Project Users <[email protected]> > Sent: Saturday, 22 August, 2009 12:53:55 AM > Subject: Re: Debug Performance in AR 2.0 > > I used dotTrace to profile it. That's how I figured out that the > debugger was the problem in the first place. When running it with the > profiler, which of course attaches to the actual exe, it runs fine and > zippy. Any recommendations on the best way to "profile" while running > in debug mode? > > As far as more AR details, we are using a single SessionScope that is > created upon application startup. Because of our key generation setup, > we are using FlushAction.Never and then manually flushing when > necessary. My investigation into the problem so far has centered > around selects; I have not looked into CUD at all. Our model is pretty > much entirely lazy-loaded. I have looked the sql nhibernate is > generating between the old and new version, and the problem is not > there. I also will tell you that, while our app is designed to be > capable of running on multiple tiers, we pulled back from WCF for > various reasons and are in fact just running multiple assemblies that > run on the local computer, attaching to an Oracle 9i database for > persistence. Let me know what other details would be helpful. > > We do have over 100 DTOs in our object model, though the selects that > are causing the problem are not reaching anywhere near that number. > > scott > > On Aug 21, 10:04 am, Krzysztof Koźmic <[email protected]> > wrote: > > > Please provide more details. Profile the application if you can. > > There's a bug in BCL that manifests itself with Dynamic Proxy when you > > create many (more than 100) proxy types. > > > Krzysztof > > > scottieb wrote: > > > Just moved our winforms, n-tier application from AR 1.03 to 2.0 and I > > > noticed the performance being 4-5 times slower for some of the gets. > > > After some investigation, I found that the problem only manifests > > > itself when running the app in debug mode from within Visual Studio; > > > running the debug exe directly works fine. Has anyone else experienced > > > this with the new version? Any ideas on what could be going on? > > > __________________________________________________________________________________ > Find local businesses and services in your area with Yahoo!7 Local. > Get started:http://local.yahoo.com.au > > WithDebugger.PNG > 150KViewDownload > > NoDebugger.PNG > 150KViewDownload --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Castle Project Users" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/castle-project-users?hl=en -~----------~----~----~----~------~----~------~--~---
